Preclinical studies have recently shown that albumin has neuroprotective effects for stroke in animal models. Thus, we sought to evaluate the effects of albumin therapy in patients with acute cerebral infarcts. We prospectively studied 49 patients with moderate-to-severe cerebral infarcts within the middle cerebral arterial territory into one of two groups: the control group (N=18) received saline, whereas the albumin group (N=31) received either 40 g or 80 g of albumin within 24 h from symptom onset. The National Institutes of Health Stroke Scale (NIHSS) and diffusion-weighted imaging (DWI) were serially checked. There was no adverse effect related to albumin therapy. Although there was no significant difference in both baseline NIHSS score and DWI lesion volume on admission, the NIHSS scores at the 14th day after treatment and the increase in DWI lesion volume 72?96 h after treatment were significantly reduced in patients of the albumin group (P=0.001 and 0.012, respectively); these effects were dose- and time-related. The outcome on the 90th day after stroke onset was more favorable in the albumin group than in the control group. However, among the albumin group, significant improvement was nearly exclusively observed in patients who had patent or recanalized vessels (P=0.004). Our results indicate that albumin therapy is a safe and effective modality in patients with acute cerebral infarction. This study also suggests that the effects of albumin therapy may vary depending on vessel status of the patient.